Status Update
No update yet.
Comments
ch...@google.com
The CL in range is aosp/3341331 and the benchmark is runBenchmark_100_RemoveMiddle_Index
We should consider removing this benchmark as it affected by unrelated changes in the code which appears to change how much code is executed making this an unreliable bencharm.
an...@google.com
Took a look here. The work we're doing to compose seems to be the same, but the new method trace now includes the measurement pass, which it didn't used to. That seems to explain this difference here, but it wouldn't explain other variability we've seen on this test, like Romain's optimizations causing significant improvements here.
I've attached the traces I captured here. Not sure what action we want to take. I suppose we could update the composition so that it doesn't have anything that remeasures?
Description
Perf Regression (High) found, matching 3 tracked metrics from benchmarks.
To triage this regression, see the guide at
Test classes affected:
Test methods affected:
Devices affected:
API Level: